Concepts & Trends

A revised chronology for the adoption of agriculture in the Southern Levant and the role of Lateglacial climatic change

A revised chronology for the adoption of agriculture in the Southern Levant and the role of Lateglacial climatic change
of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A revised chronology for the adoption of agriculture in the Southern Levant andthe role of Lateglacial climatic change S.P.E. Blockley a , * , R. Pinhasi b a Royal Holloway, University of London, Egham, Surrey TW20 0SE, UK  b Department of Archaeology, University College Cork, Cork, Ireland a r t i c l e i n f o  Article history: Received 21 January 2010Received in revised form10 September 2010Accepted 14 September 2010 a b s t r a c t This paper re-examines the chronology and environmental context for the transition to agriculture in theSouthern Levant, seen as the likely starting point for the adoption of agriculture in Europe and the NearEast. The role in this process of abrupt late Quaternary climate change has been discussed widely, butlimitations on the archaeological and palaeoenvironmental chronologies have led to varying interpre-tations. Here we attempt to clarify the situation by fi rst testing the available radiocarbon database for thearchaeological transitions from the Natu fi an through to the PPNA. We apply internationally acceptedradiocarbon quality assurance procedures and  fi nd that a signi fi cant number of the published dates fallbellow acceptable standards. The cleaning process signi fi cantly clari fi es and constrains the reported timeranges for the Natu fi an, Late Natu fi an and PPNA. We then apply the new IntCal09 calibration curve andBayesian calibration methods, using the archaeological phasing to constrain the data and calculate themost likely timing of the transitions between each phase. We then compare the onset and duration of archaeological phases to data representing the key Northern Hemisphere climatic transitions, using thenew GICC05 Greenland Ice core timescale and the timing of transitions between wet and dry phases inthe southern Levant from published high precision isotopic analyses of Speleothem data. The results of this exercise present the currently best available chronology for these events and suggest that during thesecond part of the Lateglacial interstadial, drying of the southern Levant may have triggered the tran-sition to the Late Natu fi an, when hunter-gatherer communities resorted to a more mobile lifestyle. TheLate Natu fi an culture appears to have disappeared from the southern Levant during the Younger Dryas,as drying intensi fi ed. There is then a gap in well dated evidence for human occupation until a reap-pearance of humans at the onset of the Pre-Pottery Neolithic A (PPNA) period at the beginning of theHolocene. Thus the onset of the Holocene can be hypothesised to be the driver behind the onset of theNeolithic in this region.   2010 Elsevier Ltd. All rights reserved. 1. Background and research strategy  The  “ Levantine Corridor ”  is the narrow geographic regionbetween the northeast Mediterranean Sea and the deserts to thesoutheast. During the early Holocene this corridor was very rich invegetal and animal resources and extended further east into the Jordan Valley and neighbouring regions which are now unsuitablefor rain based agriculture. According to Bar-Yosef and Belfer-Cohen(1989),Bar-Yosef(1998) theunique ecologicalandsocialconditionsthat prevailed in this southern Levantine corridor facilitated therise of early farming communities and the development of largeregional centres such as at Jericho during the PPNA period (Kuijtand Goring-Morris, 2002). More recent developments and partic-ularly work by Ofer Bar-Yosef et al. (Bar-Yosef and Belfer-Cohen,1989; Bar-Yosef and Valla, 1990) suggest a close chronologicalrelationship between the transition to agriculture and climaticdownturn during the latter part of the last glacial period. Speci fi -cally it hasbeen suggestedthatthere is alink betweencooling(andin the Levant d drying) correlated to the North Atlantic and Euro-pean Younger Dryas cold event and adaptations by Late Natu fi anhunter-gatherers (e.g., Bar-Yosef, 2001; Bar-Yosef and Belfer-Cohen,2002). While the data is complex it has been considered possiblethat the Younger Dryas induced a decrease in occupation intensityand an increase in mobility during the Late Natu fi an, and thatadoption of a subsistence practice, which predominantly relies oncultivated (wild) crops,began during the latter partof this Natu fi an *  Corresponding author. E-mail address: (S.P.E. Blockley). Contents lists available at ScienceDirect Quaternary Science Reviews journal homepage: 0277-3791/$  e  see front matter    2010 Elsevier Ltd. All rights reserved.doi:10.1016/j.quascirev.2010.09.021 Quaternary Science Reviews 30 (2011) 98 e 108  period (Bar-Yosef and Meadow, 1995; Kuijt and Goring-Morris,2002; Munro, 2003). Clearly given ongoing interest in the in fl u-ences of climate change on human adaptation (e.g., Turney andBrown, 2007) this is a signi fi cant question for Quaternary geologyand archaeology. Hence, while early agricultural communities alsoemerged in other parts of the Near East and Anatolia, the chro-nology of this southern Levantine region is particularly importantto our understanding of the timing of the early emergence of agriculture and its possible relationship with major climatic  fl uc-tuations. Addressing this problem requires us to consider twocritical issues: (a) how well can we understand the timing of climate change at a hemispherical and regional level; and (b) howreliably can we constrain the ages of the cultural transitions?Toexaminethecaseforaclimaticroleintheonsetofagriculturewe test the available chronologies for the archaeological andclimatic transitions at the end of the last glacial in the SouthernLevant. Recent reviews of the available palaeoenvironmental datafor the region suggest that the major climatic phases seen in theNorth Atlantic records are re fl ected in part in the environment of the Levant (Bar-Matthews et al.,1999, 2000; Robinson et al., 2006).Nevertheless critical issues of resolution and chronological uncer-tainty remain in both the environmental and archaeological data.For the palaeoenvironmental element to this question we incor-porate the chronologically best available hemispherical and localclimatearchives, in the form of isotopic data fromspeleothems andice. From an archaeological perspective it is necessary to test theavailable radiocarbon-based chronologies. While there is a signi fi -cant radiocarbon database for the archaeological record in theSouthern Levant this is of very mixed quality. Overthe last 50 yearsradiocarbon dating has improved signi fi cantly in quality and reli-ability. Nevertheless, only in the last decade or so has it becomeapparent that a signi fi cant number of radiocarbon ages for impor-tant time periods are either of too low resolution to be of muchpractical use, or are likely to be simply incorrect, due either tosample selection or inadequate removal of contamination. Recentdevelopments in radiocarbon dating and a growing understandingof the requirements for  “ quality assurance ”  in radiocarbon datingmean that a thorough investigation of this  14 C database is nowrequired. For example there are some older radiocarbon dates oncharcoal that when calibrated have 2 sigma uncertainties of 2000 e 3000 years, while other more precise ages are from materialthat is now considered to be unreliable, for example dates oncharred bone ordatesonlong lived woodspecies(e.g., Santosetal.,2001; Mithen and Finlayson, 2007). This important process of investigating radiocarbon databases has the potential to clarify ourunderstandingof the timingof human transitions and is nowbeingundertaken on a number of key periods in human history (e.g.,Housley et al.,1997; Blockley et al., 2006; Jacobi and Higham, 2009(European Upper Palaeolithic); Higham et al., 2009 (Neanderthal toModern Human transition); Dee et al., 2009 (Old and MiddleKingdom Egypt)). Given the importance of the transition to agri-culture in the Southern Levant it is now timely that a reappraisal of the radiocarbon database is undertaken. While some authors havediscussed quality assurance issues at speci fi c site (e.g., Mithen andFinlayson,2007)thisisthe fi rsttimethatasystematicreviewoftheradiocarbon database for transition to agriculture in the SouthernLevant haseverbeen undertaken.Additionally, whilein some casesradiocarbon dates have been rejected in the past on grounds of anexpectedageforagivenarchaeologicalculture(Hedgesetal.,1990),this is, however, subjective. Here we only apply rejection criteriabased on internationally accepted and published criteria and wehave no  a priori  assumption over what the  fi nal ages for thedifferent cultural phases should be.In order to test the database we apply (a) a new internationallyaccepted radiocarbon calibration curve this time period (IntCal09;Reimer et al., 2009); (b) a detailed analysis of the available radio-carbon information in light of issues of   “ quality assurance ”  in largeradiocarbon data-sets (see e.g., Lowe and Walker, 2000; Boarettoet al., 2003; Blockley et al., 2006); and (c) compare the latestavailable data for thetimingof climatictransitions at the end of thelast glacial, based on annual layer counting of laminated ice corerecords, using the GICC05 chronology (Rasmussen et al., 2006), andU-series dated Speleothem records, the best dated terrestrialarchives of climatic transitions in the Southern Levant (e.g., Bar-Matthews et al., 1999, 2000, 2003). The latter point is particularlyimportant, as the transition to agriculture is often discussed withreference to North Atlantic and European environmental zones,such as the Younger Dryas (e.g., Bar-Yosef and Meadow, 1995).However, only by using objective criteria for testing the availablearchaeological database and then comparing this to the bestNorthernhemisphereandlocalrecordscanwetestthereliabilityof any proposed correlation between cultural transitions and climaticchanges. 2. Archaeological context ThesouthernLevant(IsraelandJordan)ispartof thecoreregionin which agricultural communities  fi rst emerged. This has beenviewed as a three-stage process involving (a) the adoptionof a fullysedentary lifestyle by Natu fi an hunter-gatherers, dated to w 17,000 e 12,000 BP (note BP denotes calibrated radiocarbon age),followed by (b) the intensive cultivation of cereals and legumesduringthePre-PotteryNeolithicA(PPNA)period,( w 12,000 e 11,000BP.),(c)ashortEarlyPPNBperiod(11,000 e 10,400BP)(d)theMiddlePPNB,(10,400 e 9500BP),wheresubsistenceincludeddomesticatedlivestock (Bar-Yosef,1998; Kuijt and Goring-Morris, 2002).In the particular case of the southern Levant the overwhelmingevidence for the emergence of village societies and the onset of the neolithisation process are not botanical but rather focus onother archaeological criteria. Plant and animal domestication isbut one element of a complex transformation in which entailedthe  “ domestication ”  occurred in a variety of realms other thansubsistence. These include the transformation of the local envi-ronment, changes in social relations, technology, ideology, archi-tecture, and other attributes (Belfer-Cohen and Goring-Morris,2009). During the PPNA, settlement pattern contracted in scopeand shifted in location from the coastal Mediterranean belt to thenarrower strip of the Jordan Valley. The archaeological recordprovides clear evidence of a signi fi cant increase in (a) site size, (b)intensity of occupation, (c) the energetic and cultural investmentin architectural features, (d) storage facilities and (e) burial prac-tices (Kuijt and Goring-Morris, 2002; Munro, 2003; Kuijt, 2009).The PPNA subsistence spectrum combines intensive horticulture of cereals, hunting of large mammals (predominantly gazelle, and anincrease in reliance on small game such as hares, birds andreptiles). It is a subsistence strategy which is catered to theabundance of wild cereals and it is the combination of theseaspects, and particularly the emergence of new social realm whichdemarcate the onset of the transition to the Neolithic period inthis region.In this paper, we reappraise the timing of the Early and LateNatu fi an and the onset of the PPNA in the Southern Levant inrelation to the Lateglacial interstadial, the Younger Dryas and theonset of the Holocene. Fig. 1 indicates the regional study area andkeysitesthatyieldedreliableradiocarbondates(seebelow)andthemain local palaeoenvironmental record used in this study, Soreqcave (Bar-Matthews et al., 1999), as well as other Speleothemrecords discussed in the text. What is perhaps more important atthis stage is that this study will allow us to show how reliably thisquestion can be addressed given the currently available evidence. S.P.E. Blockley, R. Pinhasi / Quaternary Science Reviews 30 (2011) 98 e 108  99  3. Radiocarbon dating during the glacial-to-Holocenetransition Radiocarbondatingunderpinsallarchaeologicalchronologiesinthis time period and a very large data set of radiocarbon ages hasbeen developed for this area since the inception of the method.However, radiocarbon dating has undergone a number of criticaldevelopments over the last 50 years, and there are fundamentalissues of reliability that need to be addressed in this study. Theseproblems have been considered elsewhere and efforts have beenmade to develop objective criteria for building robust chronologies(e.g., Housley et al., 1997; Pettitt et al., 2003; Blockley et al., 2006;Higham et al., 2006, 2009). Key criteria are as follows:1) Good security of association between sample and the archae-ological event (issues such as re-working or the dating of longlived species event need to be considered; Mithen andFinlayson, 2007).2) The depositional context is understood and any localisedproblems such as  14 C reservoirs are corrected for (Shore et al.,1995).3) Suf  fi cient radiocarbon laboratory protocols are undertaken toremove contamination from the sample and to test for thepresence of contamination (such as  d 13 C measurements). Thiscan be a particular problem with older samples and radio-carbon ages on bone collagen (especially from charred bone)prior to recent developments in ultra fi ltration may be espe-cially suspect (Brock et al., 2007a,b; Higham et al., 2006).4) An internationally agreed calibration curve is available andcalibration uncertainties are accounted for in any analyses of the data (Blockley et al., 2006).5) Any statistical manipulation of the radiocarbon data, be itformal or  ad hoc  , is appropriate, and accounts for the non-normal probability densities of calibrated radiocarbon data,and reports de fi ned uncertainty ranges (e.g., Buck et al., 1991,1992). Fig. 1.  Map of study sites remaining after the radiocarbon database cleaning exercise. S.P.E. Blockley, R. Pinhasi / Quaternary Science Reviews 30 (2011) 98 e 108 100   3.1. Quality assurance Here we apply the above criteria to the analyses of large data-base of   14 C ages available for the period fromthe transition intotheLate Natu fi an through to the PPNA in Israel, Jordan, and Sinai inEgypt (the archaeological criteria for choosing these time periodsand regions are set out above). The initial database (, Pinhasi et al., 2005) consisted of 228 radiocarbon measurements on archaeological samples through tothe end of the Pre-Pottery Neolithic, with 153 of these relating toNatu fi an or PPNA samples. These dates were, however, of varyingquality, ranging from recent AMS determinations on linens andidenti fi ed plant macrofossils, or identi fi ed charcoal, through to1960 ’ s and 1970 ’ s conventional beta counting ages on unidenti fi edmaterial with exceptionally large measurement uncertainties. Inorder to  fi rst examine the range produced by these data we cali-brated all dates, including ages that had been rejected by previousauthors on the grounds that they did not  fi t with a preconceivedidea of what the timing of a particular culture should be. Theseresults are presented in Fig. 2 and show that without any qualityassurance control there is a signi fi cant spread of dates. For the LateNatu fi an in particular the range covers a period from w 30,000 to3000 BP. While the most outlying dates would have been rejectedas being  ‘ too old ’  or  ‘ too young ’  this is subjective and the problemsthat affect those dates most can also affect ages with less obviousproblems.We then attempted to apply the criteria set out above in points1 e 3 to re fi ne these data. We speci fi cally rejected dates on thefollowing grounds:1. Dates where there was not secure association in the literatureto the event in question,2. Dates where we could not ascertain what material had beendirectly dated, this is very important given the well knownproblems with radiocarbon dates on mixed material.3. Dates where we could not ascertain a likely reliable pre-treatment strategy, such as older dates on bone, wheresuccessful collagen extraction and removal of contaminantswasunlikely,againthisisasigni fi cantproblemandthedatesof poorly treated bone have been shown to be incorrect on theorder of several thousand years ( Jacobi and Higham, 2008).4. Dates that were undertaken before the late 1980s where largelaboratory measurement uncertainties meant that the ageswere likely to cause more confusion than add value (note mostof these dates were already rejected by other speci fi c qualityassurance criteria).If we examine the impact of this  fi ltering process on the datesoutlined in Fig. 2 it is clear that fi rstly the removal of old dateswithvery large errors improves the resolution of the overall chronology.Moreover these dates would have also been rejected by othercriteria. It is not possible to exactly determine how the  fi lteringprocesschangesthedatabaseinasequentialmannerasmanydatesare rejectable on several grounds. It is, however, worth noting thatour fi lteringhashadthemostimpactonthetimingoftheonsetandend of the Late Natu fi an and the onset of the PPNA. Some dates fortheLateNatu fi an,basedmainlyonburntbone,wouldplacetheendof the Natu fi an somewhere in the early Holocene, however, bonedates that have not been carefully screened and, particularly burntbone have been shown to produce unreliable ages and are thesubject of much debate (Brock et al., 2007a,b; Higham et al., 2006).Althoughwedo not have aprior viewonwhen these events shouldoccur this is clearly critical and we discuss these sites and thegeneral quality of the whole data set in more detail below.For the PPNA several of the rejected dates fall into the YoungerDryas, a keysite being the PPNA levels at Wadi Faynan (Mithen andFinlayson, 2007). Here a number of   14 C determinations wereundertaken on range of materials, using modern pre-treatmenttechniques. Despite this a range of dates were returned for twodifferent trenches. The scatter of dates ranged from  w 4800 to w 17,000 BP. There were, however, distinct populations. Dates onsnail shell returned ages of between  w 13,800 to 17,000 BP, andwere signi fi cantly older than all other methods. This was coupledwith anomalous  v 13 C values of around   5%, suggesting contami-nationoftheshellsbydeadcarbonfromthesurroundingcarbonategeology. A second set of dates on charcoal from wood that waseither unidenti fi ed or from wood identi fi ed to long lived Juniperussp,andthissetalsoreturneddatesthatweregreaterthan11,800BP. Fig. 2.  All Natu fi an and PPNA dates prior to any database cleaning, calibrated using IntCal09. S.P.E. Blockley, R. Pinhasi / Quaternary Science Reviews 30 (2011) 98 e 108  101  These dates were all several hundred years older than dates oncharcoal identi fi ed to shorter lived tree species, such as tamarix, orfromcharredseedsfromthesamecontext,whichweincludeinourcleaned database as they  fi t accepted quality assurance protocols.As there is known to be up to a 1000 year  ‘ old wood ’  effect withdating such species the dating at the site for the PPNA was con-strained only to those charcoal ages from short lived species. Thesedate the site to 11,390 e 12,240 BP.Itisworthnotingthatinordertobalancetheabilitytodiscusstheradiocarbon data with the need for a reliable database we did notapply certain additional criteria that could be reasonably imposed.Forexamplewecouldhaverejectedanydatewherea v 13 Cvaluewasnot reported in the publication. Some radiocarbon laboratoriesreport this value as a quality assurance step and fractionationcorrection while others do not, and in some cases while these aresupplied bythe laboratorytheyarenotalways reported byauthors.Hadweimposedthisconstraint,however,only39radiocarbondateswould have survived in the database. This is a key issue and thesewider chronological problems are discussed in the  fi nal section.The  fi nal result of the database cleaning exercise leaves us with w 86 out of 152 dates with good stratigraphical association andmostlybasedonAMSandconventionalmeasurementsoncharcoal,where possible identi fi ed to short lived species, and also includedare direct dates on identi fi ed wood, seeds and linen. Very few boneages are included in the remaining database, which largely re fl ectsthe dif  fi culty in obtaining bone ages in lower latitudes. The data-base of both acceptable radiocarbon dates and the rejected datesfromour study is available as Supplementary material. Fig. 3 shows the calibrated radiocarbon ages for the accepted dates prior to anystatistical analyses.  3.2. Radiocarbon calibration The second part of the criteria for building the most robustpossible assessment of a radiocarbon-based chronology iscalibration and statistical analyses. Recent improvements in theinternational effort to produce coherent radiocarbon calibrationcurves have led to the publication of IntCal09 (Reimer et al., 2009).This curve covers the period of study and is based on dendro-chronologically derived calibration data to 12,400 cal BP andmarine varve and Uranium Series dated corals for calibration databackto w 45,000 calBP. Themarine-based curvehas beenreservoircorrected to produce a terrestrial curve. Importantly there is farmore variabilityinradiocarbonproductionthantheprevious IntCal98 curve suggested, and there have been improvements in dealingwith changes in the marine reservoir offset, particularly duringcooling events such as the onset of the Younger Dryas, where thereare differences of hundreds of years between IntCal09 and oldercalibration curves (Reimer et al., 2009). It is important to test theimpact of these calibration data on the pattern of sites during theNatu fi antoPPNAtransitionaspreviousstudieshaveusedolderandless reliable curves.  3.3. Age modelling  One of the signi fi cant problems with dealing with calibratedradiocarbon dates is the need to analyse the data to test forunderlying patterns. This is compounded by the need to useappropriate methods for dealing with the often large uncertaintyranges and non-normal, multi-modal probability densities of cali-brated radiocarbon ages. Fortunately there has been signi fi cantdebate on this topic and a number of Bayesian and classicalmethods have been proposed for studying radiocarbon data andusing them to look at presence or absence of archaeological sites(e.g., Housleyet al.,1997; Blockleyet al., 2000, 2006; Blackwell andBuck, 2003). One method takes all the available calibrated radio-carbon probability densities for a particular period and these aresummed and used to identify peaks and troughs in density. Anassumption is then made that a period with greater probabilitydensity corresponds to a period of a greater number of dates and Fig. 3.  Late Natu fi an and PPNA dates for the Southern Levant after cleaning using the criteria set out in Section 3.1. The NGRIP record for the Lateglacial and early Holocene are shown for reference. S.P.E. Blockley, R. Pinhasi / Quaternary Science Reviews 30 (2011) 98 e 108 102
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks